Group numbers into classes

Recall: The transformation Group numbers into classes is possible for the open numerical variables.

This assistant allows to assign each observation for one class, that we will define, The aim is to be able later to use these classes of observations for the analyses.

In our example, we want to create classes of age. For this, we select the variable Age (numerical) then arrive to this assistant which proposes two ways of dsefining classes. Either we define the limits of classes manually (bounds of personalized classes), or we use another method allowing a distribution of observations in the classes, in function of a fixed number of classes.

Define the limits of classes

We define the limits of classes by separating them with « ; ». In our case, we wish to create the classes under 18 years, From 18 to 25 years, From 26 to 50 years and more than 50 years.

We can later modify the title of classes, created by default, by clicking , for example we can make the decision to name the under 18 years : « Minors », those between 18 and 25 years : « Active youth » etc…

• For the stages 1 and 2, to relate to the introductory sectcion of Transforming a variable.
• 3 Choose later if you wish to modify the variable chosen by recoding it (you will lose the initial variable) or if  you want to create a new variable,
• 4 In the case of the creation of a new variable, inform at minimum the name of this variable (and possibly the title),
• 5 Then choose the location of inserting a new variable created (after the variable to transform, at the end of the questionnaire).

Use classes definition method

You have the choice between four methods of the definition of classes :

• Same amplitude : The gap between the classes created will be constant. For example, in our case we will obtain the classes 20-29, 30-39, 40-49 etc…
• Same value : will create as many classes as it exists different values of the variable selected. In our example, we have chosen the variable Age, if we have three observations which the values of this variable are respectively 21, 21 et 42, the method will create two classes : the class 21, which will regroup two observations, and the class 42 with only one observation.
• Around the average : allows you to create 3, 5 or 7 classes of intervals, each interval having a « length » of ½, 1 or twice the standard deviation. The median class will be the interval containing the average. For example, your data set possesses for the variable « Age » the values 10, 20, 30, 40. We obtain an average of 25 and a standard deviation of 12,91. You wish to create five classes around the average of 1 standard deviation (12,91). These are the classes that will be created :
   
• Same (effective) : will create classes containing approximately the same number of observations (balanced distribution of observations).